Isolated population helps tease out a third locus underlying a multigenic form of canine RPGRIP1 cone-rod dystrophy

Abstract

Purpose : Canine cone-rod dystrophy 1 (cord1) is the only large animal model of inherited retinal disease where concurrent involvement of more than one gene has been identified. RPGRIP1 was initially thought to be the sole cause of cord1 while MAP9 later emerged as its age of onset modifier. Both of these gene products localize to the connecting cilia of rods and cones. In a canine research colony, we recently reported significant variability in cone ERG among RPGRIP1 mutants that could not be explained by MAP9 or environmental factors. We therefore carried out a genome-wide search for additional loci underlying the cone phenotype.

Methods : 42 dogs were genotyped using the 210K Canine chip (Illumina). All the RPGRIP1 mutants underwent complete scotopic and photopic ERGs. Genotyping data was pre-processed using Plink v1.9, and analyzed using GenABEL (GWAS), Merlin (linkage) and BEAGLE3.0 (haplotype). Retinal RNA-seq data from selected RPGRIP1 mutants with normal vs absent cone ERGs were utilized. For retinally transcribed genes in the mapped interval, 1) reads were aligned to the reference, variants were called using the bcftools/mpileup pipeline, and mutation impacts were predicted by SnpEff; 2) changes in exons were examined using IGV.

Results : GWAS of RPGRIP1 mutants with variable cone ERG revealed a single, strong association with the cone phenotype at canine chromosome 30 (-logP=7.46), and linkage suggested association in the same region. Subsequent haplotype analysis confirmed a 4Mb interval of homozygosity in 18/21 (86%) of the absent/severely-reduced cone ERG group, but only in 1/19 (5%) of the normal/mildly-reduced cone ERG group. This interval contained 49 genes of which 38 were transcribed in canine retina, and 5 missense and 14 UTR variants were found. Initial quantitative and qualitative survey of the exons found no significant changes.

Conclusions : Three recessive loci appear to concurrently contribute to the variable expression of cord1. Each locus was identified by genome-wide mapping as different canine subpopulations were developed and subsequently found to have new phenotypic variabilities. The complex hierarchical architecture e.g. epistatic vs modifier of these loci remains to be determined functionally. As the newly identified locus is directly associated with cone function, given the cone-led phenotype of cord1, a primary role in disease pathogenesis is indicated.

This is an abstract that was submitted for the 2018 ARVO Annual Meeting, held in Honolulu, Hawaii, April 29 - May 3, 2018.